General comment

This paper is a positive contribution towards the improvement of regional CTMs as it outlines the limits of a simple method that involves modulating emissions during public holidays and paves the way for further model improvements.

Authors argue that the proposed method of handling emission reductions provides positive gains in model performance scores as well as that the method "can be considered as an easy way to significantly  improve the model prediction skills". However, these are bold conclusions that can not be easily supported based on the results presented in the preprint. E.g. it turns out that the improvements in the correlation coefficients are very inhomogeneous across Europe (at least some countries like France appear to be particularly problematic with no suggested hypothesis as to why this is the case) and benefit in terms of bias in many cases relies on future improvements in emission inventories. A more veritable outlook would be that the results presented, highlight the potential for improving modelling skill by providing valuable insights into when, where and how, simple, targeted emission modulations benefit models. However, as mentioned by the authors, further in-depth analysis will be required to evolve the method in a way that more consistent results cant be obtained both spatio-temporaly and across evaluation metrics and thus render it appealing for general use.

Specific comments

- Figures 2-5, each include two stations in the Netherlands (not the same ones in all figures). Proper names of the stations (i.e. locations) should also be included rather than the cryptic NL codes, as well as the station type (rural, urban-background etc.). But there are more reasons for concern here. Understandably, these figures can not accommodate a multitude of time series from European stations, however there is no justification as to why the Netherlands provides good enough examples for time series during the examined periods, nor how these particular stations were chosen. A more fundamental concern would be that concentrating only in the Netherlands conveys a partial outlook of the effect as far as time-series are concerned, thus also hindering better understanding of the impact of the holiday emission reductions. For example, the Netherlands is not particularly known for it's Easter time festivities (figure 4), nor the May vacation happens at the exact same time across the country, thus putting into question the usefulness and certainly complicating the interpretation of figure 5 (b) and (c). Please consider using also stations from other countries with different characteristics (geographical, cultural etc.), by also providing some justification of the selection criteria.

- Paragraph 5.2: The discussion on the Figure 7 relies heavily on the claim that the performance of the model is "very good", but no accompanying evidence is presented to support that the performance is equally good in all NO2 concentration regimes. After the 24th of December, the air quality situation apparently changes and both model (BL) and observations acquire generally higher values. Can reasons other than emissions be ruled out (e.g. meteorology) or is such an increase in concentrations expected already in the BL case due to increased (!) holiday activity? In essence, does the model perform equally well (e.g. in terms of bias) in those higher concentration conditions so that we can we reliably attribute  differences between observations and model (BL) to real emission changes? If not, a possible systematic bias in the model in these different conditions could be entangled with the holiday emission reduction effects thus challenging the presented interpretation. Please consider including a different station with no such pronounced jump in NO2 levels or support better the claim that the model performs equally well in such conditions.

Technical corrections:

- line 53: Milan

- line 254: overshoot

- Please check Grivas et al. reference, not evident where this was published.

This manuscript presents and evaluates an approach to incorporate the effect of public holidays on European air quality models. The methodology consists on a simple technic based on the scaling of primary emissions at the country level when holidays occur. Two hypotheses are tested: a first one in which the levels of emissions during holidays are assumed to be equal to the Sundays ones, and a second one where an 80% emission reduction is considered during holidays. Both approaches are evaluated by comparing modelled results against in-situ observations. As described by the authors in the introduction section, which I find it very detailed and comprehensive, several observations-based studies have already highlighted the effect that specific holidays can have on pollutant concentrations. However, a systematic analysis of the holidays effect from a modelling perspective at the EU scale, as well as the description of a methodology to properly incorporate it into AQ models has not yet been addressed with the same level of detail. Therefore, the topic presented in this manuscript is of interest and represents a good contribution to GMD. Nevertheless, there are several aspects of the manuscript – including methods, evaluation and discussion of the results – that, in my view, are not sufficiently convincing in their current form and should be carefully revised before the manuscript can be accepted for publication.

Major comments:

C1. In one of the sensitivity tests, authors assume an 80% emissions reductions during holidays. Authors already mention in the manuscript that the presented approach should be considered only as a first step. However, I think that the hypothesis made (80% reduction for all sectors considered) should be backed up by the analysis of sectoral related activity data. While it is true that for certain sectors this analysis can be challenging due to the lack of data, for others there is information that can be used for this purpose. In the case of the A_PublicPower industry sector, authors could use the ENTSO-E transparency platform (https://transparency.entsoe.eu/), which reports data on hourly electricity generation by fuel type per country. In the case of F_RoadTransport, authors could use information on traffic counts reported by national transport agencies, such as the Finish transport agency (https://vayla.fi/en/transport-network/data/open-data/road-network/tms-data).

C2. Following with the previous comment, it is questionable that all emissions from the C_OtherStationaryCombustion sector suffer an 80% reduction during holidays. In the case of PM, between 80 and 90% of total emissions are related to residential wood combustion activities (https://www.ceip.at/). Several studies have shown that residential wood combustion activities in Europe tend to increase significantly during weekends when compared to weekdays, as people use this fuel for recreational purposes. Examples of these studies are Krecl et al. (2008) and Athanasopoulou et al. (2017). I would expect a similar behavior during holidays (specially Christmas period), when people spend more of their time at home. The PM10 and PM2.5 spatial scores for Christmas shown in the supplementary material already suggest that with the holiday days considered as Sundays (the HS case) the skills of the model deteriorate (correlation decreases and MB increases). Authors suggest that this could be related to the use of fireworks, which are not accounted in the CAMS emission inventory, but this is not proved in the manuscript.

C3. In the abstract section, authors mention that “Spatial and temporal distributions of atmospheric concentrations of the major air pollutants (PM2.5, PM10, SO2, CO, NO2, NOX, and O3) were considered”. However, the analysis, evaluation and discussion of the results is very much focused on NO2. Figures 2 to 7 show NO2 results, while results for CO, O3 and PM2.5 are only shown on Figures 8 and 9 (no results for SO2 or PM10 appear in the main manuscript, only in the supplementary material). A more balanced discussion of all the pollutants considered should be provided (or, alternatively, the pollutants not shown in the main manuscript could be removed from the study). In the case of O3, the discussion is focused on the Christmas period, when O3 levels are very low. Discussions for this pollutant should be focused on Easter. In the case of NO2, time series are shown almost exclusively for stations in the Netherlands (Figures 3,4,5 and 7). Considering that the study is performed at the EU level, it would be interesting to see specific results in other regions.

C4. The same prescribed sector-specific emission diurnal profiles are used for weekdays, weekends and holidays, which is a limitation of the study. Authors mention several times in the text that the incorporation of specific weekend and holidays diurnal profiles should be done when available. However, several works have already reported in the past specific Saturday/Sunday diurnal profiles for the road transport sector, which is the main contributor to total NOx emissions. Examples of these profiles can be found in Pregger et al., (2007); Menut et al. (2012); Mues et al. (2014) and Guevara et al. (2021), among others. Following with the hypothesis made by the authors at the weekly level, the same diurnal profiles proposed for Sundays could be assumed for holidays, at least for the road transport sector. I think this point should be addressed more carefully, and perhaps it would be good to produce an extra AQ run assuming a set of specific weekend/holidays diurnal profiles. This would, for sure, bring an added value to the study.

Other comments:

1. In section 2.1 authors mention that they consider events marked with “National holiday”, “local holiday” and “common local holiday” when retrieving holiday events from the Calendarific API. I understand that “local holiday” and “common local holiday” refer to holidays that are only occurring in a specific region(s) of the country – while in the rest of the country is a normal working day. Considering that the emission scaling approach proposed is at the country-level, should not only “National holiday” be considered?
2. In Table 4 – Meteorological driver, should not be “interpolated to 0.2x0.2” instead of “0.1x0.1”?
3. Line 135, define GNFR acronym + revise the number of GNFR sector (it is 16 and not 7. Note that the GNFR_F sector is split by fuel type)

References:

Pregger, T., Scholz, Y. & Friedrich, R. Documentation of the Anthropogenic GHG Emission Data for Europe Provided in the Frame of CarboEurope GHG and CarboEurope IP - Final Report. Stuttgart, http://carboeurope.org/ceip/products/files/Pregger_IER_Final_Report_Feb2007.pdf (2007).

Menut, L., Goussebaile, A., Bessagnet, B., Khvorostiyanov, D., and Ung, A.: Impact of realistic hourly emissions profiles on air pollutants concentrations modelled with CHIMERE, Atmos. Environ., 49, 233–244, https://doi.org/10.1016/j.atmosenv.2011.11.057, 2012.

Mues, A., Kuenen, J., Hendriks, C., Manders, A., Segers, A., Scholz, Y., Hueglin, C., Builtjes, P., and Schaap, M.: Sensitivity of air pollution simulations with LOTOS-EUROS to the temporal distribution of anthropogenic emissions, Atmos. Chem. Phys., 14, 939–955, https://doi.org/10.5194/acp-14-939-2014, 2014. 

Guevara, M., Jorba, O., Tena, C., Denier van der Gon, H., Kuenen, J., Elguindi, N., Darras, S., Granier, C., and Pérez García-Pando, C.: Copernicus Atmosphere Monitoring Service TEMPOral profiles (CAMS-TEMPO): global and European emission temporal profile maps for atmospheric chemistry modelling, Earth Syst. Sci. Data, 13, 367–404, https://doi.org/10.5194/essd-13-367-2021, 2021.

Krecl, P., Hedberg Larsson, E., Ström, J., and Johansson, C.: Contribution of residential wood combustion and other sources to hourly winter aerosol in Northern Sweden determined by positive matrix factorization, Atmos. Chem. Phys., 8, 3639–3653, https://doi.org/10.5194/acp-8-3639-2008, 2008.

Athanasopoulou, E., Speyer, O., Brunner, D., Vogel, H., Vogel, B., Mihalopoulos, N., and Gerasopoulos, E.: Changes in domestic heating fuel use in Greece: effects on atmospheric chemistry and radiation, Atmos. Chem. Phys., 17, 10597–10618, https://doi.org/10.5194/acp-17-10597-2017, 2017.